The quality of groundwater resources in artisanal mining districts in Ghana is under threat due to pollution; rendering the resource unsafe for drinking and irrigation purposes. This makes the assessment of the quality of groundwater resources a relevant aspect of groundwater studies as it informs decision making and monitoring. This study adopts 3 Machine Learning (ML) models, Support Vector Regression (SVR), Gradient Boost Regression (GBR), and Artificial Neural Network (ANN), to evaluate a variety of irrigation water quality metrics such as Sodium Percentage (Na%), Soluble Sodium Percentage (SSP), Sodium Adsorption Ratio (SAR), Residual Sodium Carbonate (RSC), Permeability Index (PI), Pollution Index of Groundwater (PIG), Kelly’s Ratio (KR), and Magnesium Hazard (MH). 105 samples were collected from a mining area in Northern Ghana and analysed through traditional methods. The Irrigation Water Quality Indices (IWQIs) demonstrate that all water samples are suitable for use as irrigable water with the exception of MH, Na%, PI, and PIG which revealed that 69.52%, 8.57%, 29.52%, and 3.81% are inappropriate for irrigation. SVR, GBR and ANN were used to establish important factors that may influence IWQIs in the area. The measured data was used as independent variables, and the derived IWQIs, the dependent variables. The results revealed that ANN, GBR, and SVR are all viable options for the prediction of IWQIs, but GBR exhibited variable performance in some indices making it lack consistency and thus falls a bit short compared to ANN and SVR. SVR models overall performed best with SVR-RSC having the highest accuracy.